agentleFS
JakeschincariolJakeschincariol/seed-skill/seed-updates/SKILL.md

seed-updates

Skill51 starsChanged 11 days agoMITView on GitHub

Investor Updates. Writes the monthly investor update that keeps investors warm: a TL;DR, every metric with its month-over-month growth, highlights, lowlights, the asks, and the runway worked out from your cash and your average net burn. A short "prospects" version goes to warm investors who said "not now": the headline numbers, two highlights, the raise and an opt-out line, with no cash or runway. Reads seed/metrics.json (and seed/config.json for the founder and the raise), writes seed/updates/YYYY-MM.md. Never sends anything on its own. Use when the user says "write my investor update", "monthly update", "update my investors", "keep investors warm", "follow up with investors who said not now", "what's my runway", "MoM growth", or at the start of each month.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run update.py write --month YYYY-MM.

    line 83
  2. Distinctive:

    Never send on the founder's behalf without a yes for that batch, and only through their own account.

    line 125

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

seed-updates

Investor Updates is the habit that makes the next round easier. Once a month, a short, honest email: what the numbers did, what went well, what did not, what you need, and how long the money lasts. Investors who already backed you can help when they know the asks; investors who said "not now" watch the line go up and come back. This skill keeps the numbers in one file, works out the growth and the runway exactly, and writes the update in the standard shape. It writes. It does not send: the founder reads it and says yes first.

${CLAUDE_SKILL_DIR} below is this skill's folder (where this SKILL.md lives).

python3 "${CLAUDE_SKILL_DIR}/update.py" init                                 # seed/metrics.json from the example
python3 "${CLAUDE_SKILL_DIR}/update.py" add --month 2026-12 --set revenue=7875 --set walks=2100 --cash 160000
python3 "${CLAUDE_SKILL_DIR}/update.py" add --month 2026-12 --highlight "..." --lowlight "..." --ask "..."
python3 "${CLAUDE_SKILL_DIR}/update.py" write --month 2026-11                # seed/updates/2026-11.md
python3 "${CLAUDE_SKILL_DIR}/update.py" write --month 2026-11 --to prospects # seed/updates/2026-11-prospects.md
python3 "${CLAUDE_SKILL_DIR}/update.py" write --stdout                       # the latest month, printed only

Other flags: --dir PATH (default ./seed), add --money-in N (funding that arrived that month, which is not burn), add --net-burn N (override), add --tldr TEXT (your own TL;DR line), init --force.

What it reads and writes

filewhat
metrics.example.json (this folder)the Leashly pilot example init copies
seed/metrics.jsonyour metrics (key, label, unit: usd, count or pct), the headline three, and one entry a month: values, cash at month end, money_in, highlights, lowlights, asks, optional tldr
seed/config.jsonfrom /seed-plan: the company name, the founder's name for the sign-off, the raise for the prospects version
seed/updates/YYYY-MM.mdwrite: the investor update
seed/updates/YYYY-MM-prospects.mdwrite --to prospects: the short version for "not now" investors

The method

  • Growth: (this month minus last month) / last month, one decimal, for money and counts ("up 23.9% from $5,325"). Percent metrics change in points ("down 0.4 pts from 2.8%"). A metric with no last month says "new this month".
  • Net burn: last month's cash + money that came in this month minus this month's cash. A raise landing is not negative burn: record it as money_in.
  • Runway: this month's cash / the average net burn of the last 3 months (fewer if that is all there is), one decimal. "To around" = this month plus the runway, rounded up to a whole month. If the cash is not running down, it says so instead of a number.
  • TL;DR: your own line if the month has tldr, otherwise the three headline metrics with their growth and the runway.
  • Honest by default: a month with no lowlights, no highlights or no asks is flagged. Lowlights build trust; asks get you help.
  • Prospects: the headline numbers and the top two highlights, the raise from config.json, "Hi {first_name}," for you to fill in per investor, and the opt-out line: "Not a fit any more? Reply 'no' and I won't send these again." Cash, burn, runway, lowlights and asks stay out on purpose.

Steps

  1. First time: update.py init, then replace the Leashly metrics with the five or six numbers this company tracks every month. Keep them the same month to month.
  2. Each month, ask the founder for the month-end numbers, the cash in the bank, any money that came in, and three highlights, the lowlights and up to three asks. Record them with update.py add.
  3. Run update.py write --month YYYY-MM. Read the TL;DR, the numbers and the runway back to the founder. Fix wrong numbers in seed/metrics.json and write again.
  4. For investors who replied "not now" (in /seed-closer's seed/pipeline.json, status passed with a "not now" reply), run write --to prospects. Never send it to anyone who said no, opted out or bounced.
  5. Sending: show the founder the recipients and the full email, and wait for a clear yes for that batch. Then send it through the founder's own Gmail connector (or save Gmail drafts if they prefer). Fill in {first_name} for each prospect. Nothing goes out on its own.
  6. Hand off: /seed-closer logs any reply and books a call; /seed-model updates the assumptions when real numbers drift from the plan.

Example: the reel

Leashly (fictional) runs a one-city pilot before the seed. November 2026 in seed/metrics.json: revenue $6,600, 1,760 walks, 210 active owners, $44,000 GMV, 31 walks saved by a backup walker, 2.4% walker cancellations, $177,800 cash at month end (October: $196,500; September: $214,000). update.py write --month 2026-11:

Subject: Leashly investor update: November 2026

**TL;DR:** Revenue $6,600 (+23.9% MoM), walks completed 1,760 (+23.9%), active owners 210 (+22.8%). Runway 9.8 months.

- Revenue: $6,600, up 23.9% from $5,325
- Walker cancellations: 2.4%, down 0.4 pts from 2.8%
...
Cash at the end of November: $177,800. Net burn this month: $18,700. Average net burn over
the last 2 months: $18,100 a month, so the cash lasts about 9.8 months, to around Sep 2027.

The net burn is $196,500 - $177,800 = $18,700 in November and $214,000 - $196,500 = $17,500 in October; the average is $18,100, and $177,800 / $18,100 = 9.8 months. The prospects version opens "Hi {first_name}," and ends: "We are raising a $1.5M seed (post-money SAFE). If the timing is better now, would a 20-minute call be useful?"

Rules

  • Never send on the founder's behalf without a yes for that batch, and only through their own account. Respect every "no" and opt-out forever.
  • Real numbers only. Never round a number up, drop a bad metric or skip a lowlight to make a month look better.
  • Runway is arithmetic on the founder's own cash and burn, not a promise. Not investment advice.
  • The prospects version never shares cash, burn or runway.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.